Materials Science and Engineering C 59 (2016) 754–765

Contents lists available at ScienceDirect

Materials Science and Engineering C journal homepage: www.elsevier.com/locate/msec

Ultrafine-grained porous titanium and porous titanium/magnesium composites fabricated by space holder-enabled severe plastic deformation Yuanshen Qi a,⁎, Karla G. Contreras b, Hyun-Do Jung c,d, Hyoun-Ee Kim d,e, Rimma Lapovok a, Yuri Estrin a,f,⁎⁎ a

Centre for Advanced Hybrid Materials, Department of Materials Engineering, Monash University, Clayton, Victoria 3800, Australia Monash Institute of Medical Engineering, Faculty of Engineering, Monash University, Clayton, Victoria 3800, Australia c Liquid Processing & Casting Technology R&D Group, Korea Institute of Industrial Technology, Incheon 406-840, Republic of Korea d Department of Materials Science and Engineering, Seoul National University, Seoul 151-742, Republic of Korea e Advanced Institutes of Convergence Technology, Seoul National University, Gwanggyo, Yeongtong-gu, Suwon-si, Gyeonggi-do 443-270, Republic of Korea f Laboratory of Hybrid Nanostructured Materials, NUST MISiS, Moscow 119490, Russia b

a r t i c l e

i n f o

Article history: Received 20 March 2015 Received in revised form 15 October 2015 Accepted 23 October 2015 Available online 26 October 2015 Keywords: Equal channel angular pressing compaction Space holder method Porous Ti Porous Ti/Mg composite Microstructure

a b s t r a c t Compaction of powders by equal channel angular pressing (ECAP) using a novel space holder method was employed to fabricate metallic scaffolds with tuneable porosity. Porous Ti and Ti/Mg composites with 60% and 50% percolating porosity were fabricated using powder blends with two kinds of sacrificial space holders. The high compressive strength and good ductility of porous Ti and porous Ti/Mg obtained in this way are believed to be associated with the ultrafine grain structure of the pore walls. To understand this, a detailed electron microscopy investigation was employed to analyse the interface between Ti/Ti and Ti/Mg particles, the grain structures in Ti particles and the topography of pore surfaces. It was found that using the proposed compaction method, high quality bonding between particles was obtained. Comparing with other powder metallurgy methods to fabricate Ti with an open porous structure, where thermal energy supplied by a laser beam or high temperature sintering is essential, the ECAP process conducted at a relatively low temperature of 400 °C was shown to produce unique properties. © 2015 Elsevier B.V. All rights reserved.

1. Introduction The concept of using porous structure to achieve adequate and permanent fixation of a bone implant to the skeletal system was first investigated in the 1970s [1,2]. At that time, instead of open porous structures, materials with a porous surface were extensively studied as a way to achieve implant fixation. Indeed, the solid inner core of an implant was considered to be necessary to provide the mechanical strength for load-bearing applications [2,3]. In vitro and in vivo tests were done for tailoring the pore size to achieve high rate of bone ingrowth [1,4]. Clinical tests showed encouraging results indicating that biological fixation occurred through bone ingrowth. However, there was also a concerning aspect regarding the bone resorption and bone loss caused by stress shielding, owing to excessively high stiffness of the rigid implant. In a clinical trial, over 12% of patients suffered from moderate or severe bone resorption within two years after the implantation [5]. Due to

⁎ Corresponding author. ⁎⁎ Corresponding author at: Centre for Advanced Hybrid Materials, Department of Materials Engineering, Monash University, Clayton, Victoria 3800, Australia. E-mail addresses: [email protected] (Y. Qi), [email protected] (Y. Estrin).

http://dx.doi.org/10.1016/j.msec.2015.10.070 0928-4931/© 2015 Elsevier B.V. All rights reserved.

this stress shielding, the use of large and rigid femoral implants was not recommended for patients whose bone mass was initially low [6]. To minimize the stiffness mismatch between the implant and the surrounding bone tissue and achieve optimum load transfer from the artificial implant to the adjacent bone, implants with low elastic modulus are being considered [7]. Based on this demand, new β-titanium alloy compositions with a low elastic modulus have been under development as an alternative to stainless steels, cobalt-based alloys, and α-titaniumbased alloys [8,9]. With porous surface of solid core β-Ti, good fixation of the implant is expected. However, there are still concerns related to this route because solid core material still has higher stiffness than bone. Besides, bone mass is lost and bone formation ability decreases as people age [10]. To further decrease the Young's modulus of the implant and achieve a higher degree of osseointegrated fixation, a highly open porous structure appears to be more attractive than a solid implant with porous surface [11]. Currently, highly porous Ti based materials are attracting a great deal of attention from the academic community and medical implant industry alike, as Ti and its alloys possess outstanding mechanical properties, corrosion resistance, and biocompatibility [9,12–15]. The excessively high stiffness of an implant material and the associated stress shielding effect can readily be reduced by introducing

Y. Qi et al. / Materials Science and Engineering C 59 (2016) 754–765

porosity. However, a downside of this approach is a significant loss of strength, which reduces the ability of the implant to withstand the mechanical loads placed upon it during service [16,17]. Furthermore, with insufficient stiffness, interface de-bonding and implant loosening may occur [18]. Thus, the first and biggest problem porous Tibased materials are facing today is inadequate strength. There are a number of factors that can affect the compressive strength of the porous structure, including the level of porosity, pore size, pore shape and interconnectivity of the structure [16,19,20]. To mimic the biomechanical properties of natural bone, these parameters need to be tailored to be close to those of the surrounding bone tissue. Ultimately, the strength of the structure depends on the properties of pore walls [21]. In the case of open porous Ti produced from Ti powders, bonding between the particles and their microstructure determine the properties of pore walls. In one of the current approaches, particles are bonded together by laser melting. Selective laser melting is a promising additive manufacturing technology that can be used to fabricate porous orthopaedic implants [22–24]. However, its use is limited by the lack of control of the microstructure and the surface roughness due to thermal gradients that develop during rapid solidification [25]. Porous Ti can also be manufactured by slurry foaming [26] and freeze casting [27], whereby pores are formed by decomposition of the foaming agent and evaporation of frozen solvent crystals, respectively. Another extensively studied facile approach to fabricate porous Ti is the space holder method [28]. Ample literature devoted to this method was reviewed in Ref. [29]. Although the above mentioned methods involve different routes, sintering in vacuum at around 1300 °C invariably needs to be conducted as the last processing step to enhance inter-particle bonding and increase the strength of the struts [30]. In other words, a large thermal energy is needed to either melt or vacuum sinter Ti particles to enhance diffusivity of Ti and oxygen atoms in the oxide layer thus achieving sufficient bonding between Ti particles. Not only is this heat input costly [31], but it also causes undesirable grain coarsening. By contrast, the technique we propose below does not involve the use of excessively high temperature and as such makes it possible to obtain a fine grain structure of the porous product. Indeed, grain size is a key microstructural factor that has a significant effect on nearly all aspects of the physical and mechanical behaviour of polycrystalline metals. It also affects their chemical and biochemical response to the surrounding media, such as bodily fluids and tissues [32]. Recently, ultrafine-grained (UFG) and especially nano-grained (NG) metals came to the fore in the context of implant applications as a promising alternative to alloys, which in many cases have adverse biological effects due to the release of alloying element ions [33–36]. A particular avenue to obtaining bulk UFG/NG materials is severe plastic deformation (SPD) [32,37]. Grain structure refinement and modification by SPD has been successfully used to obtain biomedical materials with improved biocompatibility and mechanical performance [38–43]. Comprehensive reviews of studies on cellular response to UFG/NG materials produced by SPD can be found in Refs. [44,45]. The SPD techniques, such as equal channel angular pressing (ECAP), which were initially developed for processing bulk billets, can also be used for powder compaction [46]. An important benefit of ECAP with back-pressure as a means of powder compaction is that it can be carried out at lower temperatures than the conventional powder consolidation techniques. Owing to severe shear deformation combined with high hydrostatic pressure, this method offers higher density, finer microstructure, and possibly stronger bonding of the compact than traditional powder metallurgy techniques do [47,48]. In this context, the experience of fabrication of multicomponent materials using ECAP compaction can be used in conjunction with the space holder method to replace the traditional powder consolidation techniques. This approach, combining the mentioned benefits of compaction of powders by ECAP with the use of sacrificial space-holder powders, was consequently followed in the present study. The novelty was the

755

use of two types of space-holder particles. Blends of commercial purity (CP) Ti powder as the base material and Mg and Si powders as the space-holders were used. CP Ti, rather than a commonly used alloy, such as Ti–6Al–4V, was chosen because it does not release toxic alloying elements in the body. Likewise, neither Mg nor Si causes cytotoxicity [49,50]. Adding Mg had three objectives. The first one was to increase the processability of the mixture during deformation. The second one was to explore the potential of a porous Ti/Mg composite. The philosophy behind this concept is a recent proposal to add Mg as a biodegradable material to Ti strut, thus providing higher mechanical strength to the implant initially, whilst enabling its load bearing capacity and stiffness to decrease gradually as Mg dissolves away giving way to the ingrowing bone tissue [51,52]. Finally, the use of Mg as space-holder in titanium makes it possible to provide the compact with machinability required for implant manufacturing. Porosity can be introduced after machining by leaching Mg particles from the manufactured implant. In what follows, we describe the procedures used to fabricate porous Ti and Ti/Mg composites with different levels of porosity using compaction of powders by ECAP, with subsequent removal of sacrificial spaceholder particles. We also present the results of mechanical testing and analysis of pore structure and the character of bonding between particles as a function of the processing conditions. Finally, cell response to the structures produced is discussed.

2. Materials and method 2.1. Preparation of Ti/Mg/Si composites by ECAP compaction CP Ti powder (Grade 2, −325 mesh, Sumitomo, Japan), Mg granules (99.6% pure, −200 mesh, Materion, USA) and Si particles (99.9% pure, −40 mesh, Hokin, China) were used as the starting materials. Scanning electron microscopy (SEM) images of the powders are shown in Fig. 1, the average particle size being 49.5 ± 9.5 μm, 89.7 ± 16.8 μm, and 415.7 ± 39.5 μm, respectively. Since the level of porosity and pore size both play a critical role in bone ingrowth, the minimum requirement for pore size for cell migration and transport is considered to be ∼100 μm; higher porosity and larger pore size result in greater bone ingrowth [19]. The composition of Ti/Mg/Si powder blend by volume ratio was chosen as Ti40Mg10Si50 (40 vol.% Ti, 10 vol.% Mg and 50 vol.% Si). 50 vol.% of Si is necessary here to ensure pore interconnectivity once this constituent is removed. The powder blends were weighted in an argon-filled glove box, then mixed and stirred with 1 wt.% of isopropyl alcohol for 3 min. Subsequently, this elemental mixture slurry was taken from glove box in a sealed container and poured into the entry channel of ECAP die, whilst the exit channel was blocked by a back-pressure punch. The mixture was cold compacted at 50 MPa and dried at room temperature for 5 min. Cold compaction as a pre-ECAP step was used to lock the position of each constituent because otherwise powder segregation occurred when the isopropyl alcohol evaporated and the adhesion force between particles disappeared. It is worth noting that good blending and avoiding segregation are essential to ensure a uniform distribution of all components in the composite and the desired interconnectivity of the porous structure. ECAP consolidation process was conducted at 400 °C. One, two and four ECAP passes were used following Route Bc, which involves rotation of the sample by 90° about the long sample axis between the passes. The detailed operation of the ECAP apparatus was described in our previous publication [53]. The equivalent true strain (ɛ) the material underwent in one pass was 1.15. With increasing number of ECAP passes a very large strain was accumulated. The cumulative equivalent strain was chosen as the main variable in this study to investigate the effect of plastic deformation of Ti/Mg/Si compacts on the mechanical properties and the biocompatibility of the porous structures produced.

756

Y. Qi et al. / Materials Science and Engineering C 59 (2016) 754–765

Fig. 1. Scanning electron microscopy images of the initial powder materials: (a) titanium powder; (b) magnesium powder, and (c) silicon powder.

2.2. Machinability of Ti/Mg/Si compacts Machinability is usually an overlooked property, despite being a determining factor in fabrication of a desired shape of the implant. The importance of machinability of the compact, and the advantages of the space holder method in this regard were recently discussed by Kim et al. [54]. It was found that because Mg has higher strength and elastic modulus than NaCl as a space holder, the transverse rupture strength of Ti/Mg compact is higher than that of Ti/NaCl. This also leads to better machinability of Ti/Mg compacts. To verify the feasibility of machining the Ti40Mg10Si50 composite fabricated by ECAP compaction, turning, drilling and wire cutting were performed on the compacts. For drilling and turning, high speed steel was used as cutting tool. Meanwhile, the cutting speed, feed rate and depth of cut were chosen as 25 m/min, 0.5 mm/rev and 0.1 mm, respectively. 2.3. Synthesis of porous Ti/Mg and porous Ti The compacts produced by ECAP were soaked in aqueous 5 M sodium hydroxide (NaOH) solution at 60 °C for 12 h to remove Si, washed and ultrasonically cleaned with distilled warm water. This recipe is similar to the one used as alkali treatment to make Ti surface bioactive for apatite precipitation by forming a thin sodium titanate layer [55]. To remove magnesium — the second sacrificial ingredient, porous Ti/Mg composites were immersed in 100 mM hydrochloric acid (HCl) at room temperature for 6 h followed by distilled warm water wash and air dry for 24 h. The reactions used to remove Si and Mg are given below: SiðsÞ þ 2NaOHðaqÞ þ H2 OðaqÞ ¼ Na2 SiO3ðaqÞ þ 2H2ðgÞ

ð1Þ

MgðsÞ þ 2HClðaqÞ ¼ MgCl2ðaqÞ þ 2H2ðgÞ

ð2Þ

The evolution of the material from Ti/Mg/Si compact to porous Ti/Mg composite with porosity of 50%, and then to porous Ti with 60% porosity can be followed in Fig. 2. The porous Ti/Mg composite seen in Fig. 2b was obtained by removing Si particles using reaction (1), whilst porous Ti depicted in Fig. 2c was fabricated as a next step, by leaching Mg out of the porous Ti/Mg composite according to reaction (2). Altogether, three types of materials (Ti/Mg/Si, Ti/Mg, and Ti) were considered — two of them (Ti/Mg and Ti) with porosity — using three ECAP histories: one, two, and four passes. Hence, six types of porous structures were investigated. 2.4. Characterization of porous Ti/Mg and porous Ti A low-magnification overview of the porous structures was obtained using Olympus SZX-16 stereo optical microscope with digital camera. The morphology of the pores was examined using JEOL 7001 field emission gun (FEG) scanning electron microscope (SEM) and FEI Quanta FEG SEM. Porosity and pore interconnectivity were studied using Skycan 1173 (Kontich, Belgium) micro-tomography scanner (μ-CT). 2.5. Measurement of mechanical properties In order to evaluate the mechanical properties of porous Ti/Mg and porous Ti, uniaxial compression tests were conducted on cylindrical specimens with the height of 7 mm and diameter of 4.5 mm, which are shown in Fig. 2b and c. Compression test samples were cut along the ECAP direction by wire cutting. Compression tests were carried out using an Instron 5982 machine with a cross-head velocity of 0.007 mm/s corresponding to a nominal strain rate of 10−3 s− 1. The measurements were done after a preload of 30–50 N, which was high enough to minimize potential problems with setting of the specimens

Fig. 2. Optical micrographs of (a) the initial Ti/Mg/Si compact after compaction by a single ECAP pass, (b) porous Ti/Mg composite after Si particles were leached out, and (c) porous Ti after Mg granules were removed.

Y. Qi et al. / Materials Science and Engineering C 59 (2016) 754–765

757

Fig. 3. SEM images of (a) external thread fabricated by turning and (b) internal hollow core produced by drilling. Insets (c) and (d) are the corresponding enlarged images.

in the machine. Hence, the initial straight-line portion of the stress– strain curve within the elastic limit was measured with sufficient confidence. The elastic compliance of the Instron machine was determined by compression testing of bulk Grade 2 titanium and was considered when determining the Young's modulus of the porous specimens. 2.6. Characterization of microstructure of pore walls With increasing number of ECAP passes, a higher degree of grain refinement and better bonding were expected in pore walls in Ti. To measure the grain size and interfacial microstructure, JEOL 7001 FEG SEM equipped with OI Aztec electron backscattered diffraction (EBSD) system and 200 kV FEI Tecnai F20 FEG scanning transmission electron microscope (STEM) equipped with Bruker Quantax 400 STEM X-ray analysis system were used. EBSD samples were prepared by hand grinding/polishing followed by ion beam milling and transmission electron microscopy (TEM) samples were prepared by wedge polishing and ion beam milling. 2.7. Cell culture and cell viability Cell compatibility was investigated by in vitro test using preosteoblast cell line MC3T3-E1. Cells were cultured in α-Minimal Essential Media (Gibco, Life Technologies) supplemented with 10% Foetal Bovine Serum (Gibco, Life Technologies), 16.8 mM HEPES buffer (Gibco, Life Technologies), 1% penicillin/streptomycin (Gibco, Life Technologies) and incubated at humidified incubator at 37 °C, 5%CO2. At 70% confluence, cells were harvested and plated at a density of 1 × 104 cells per

well in a 96 multiwell plate. Porous Ti disc with 5 mm diameter and 1 mm thickness were placed in a 96 multi-well plate and seeded with 1 × 104 cells per well. Positive control wells containing dense Grade 2 Ti samples of the same size were set up in parallel. Experiment was set up in triplicate and plates were incubated at 37 °C, 5% CO2. Cell viability was determined by MTS assay (Cell Titre 96 Aqueous One Solution Cell Proliferation Assay, Promega) at day 3 and day 7 following manufacturer instructions. For SEM analysis, cells were fixed onto the porous Ti discs using 2.5% glutaraldehyde in 0.1 M sodium cacodylate buffer for 4 h and postfixed in 1% osmium tetroxide for 2 h at room temperature. The fixed samples were then dehydrated in an ethanol series followed by a hexamethyldisilazane drying procedure. Finally, all samples were air dried for 30 min, sputter-coated with gold, and characterized using a FEI Nova NanoSEM. 3. Results 3.1. Machinability Wire cut samples are shown in Fig. 2a and for turning and drilling, a hollow screw with the external thread M10 × 1 × 5 (shown in Fig. 3a) and inner diameter of 6 mm (shown in Fig. 3b) was fabricated. The hollow version mimics a cannulated screw. The figure demonstrates the integrity of the structure after machining and subsequent leaching of the sacrificial material. Moreover, with a closer inspection of Fig. 3(c,d) it can be seen that no cracks initiated from the interfaces between pores and Ti struts, which indicates good bonding at interfaces between Si and Ti/Mg particles.

Fig. 4. Micro-CT images for the porous Ti/Mg and porous Ti produced by ECAP compaction after (a,d) 1 pass, (b,e) 2 passes and (c,f) 4 passes.

758

Y. Qi et al. / Materials Science and Engineering C 59 (2016) 754–765

3.2. Porosity and interconnectivity Pore morphology in the porous materials produced is presented in Fig. 4. Cylindrical samples 12 mm in diameter and 5 mm in height were cut into two halves. One half was made to porous Ti/Mg and the other one to porous Ti by the leaching techniques described above. They were characterized by μ-CT, and the corresponding images are presented in Fig. 4. It can be seen that pores are distributed uniformly, which indicates that the mixing of the three constituent powders was homogeneous. The μ-CT data confirm 100% interconnectivity in all six samples. This full interconnectivity is a result of the volume fraction of Si particles being sufficient high. The level of porosity in Ti/Mg and Ti were calculated to be 50% and 60%, respectively. It should be mentioned that during wet mixing Ti and Mg particles adhered to Si granules and after the latter were removed, Mg particles were exposed to open air. Percolating porosity in porous Ti/Mg enables the use of this material in partly bioresorbable implants. The Mg constituent will be dissolved in bodily fluids providing the surrounding bone tissue with the ability to grow into a porous Ti scaffold that is left. The average pore size and wall thickness calculated from the μ-CT results are presented in Fig. 5. It is seen that the fourth ECAP pass gives rise to a slight increase in the average pore size and wall thickness of the porous Ti/Mg composite after two ECAP passes. In contrast, for porous Ti, an increase in the number of ECAP passes is accompanied with a substantial increase in both the pore size and the wall thickness. This may be associated with a redistribution of the constituents during ECAP. This observation will be explained in detail in Section 4.1. 3.3. Mechanical properties Fig. 6 shows representative stress–strain curves for the six kinds of porous materials investigated. Due to the elimination of sintering at 1300 °C in the processing schedule used, the oxygen content of the materials produced is believed to be low and the curves do not show brittle behaviour [54,56]. In the case of porous Ti and porous Ti/Mg whose processing history involved 4 ECAP passes, a strain hardening region is observed after plastic yielding. A reasonably good combination of compressive strength and ductility as reflected in the strain to failure was obtained, with the Ti/Mg composite that experienced 4 ECAP passes in its processing history showing the best results. The compressive strength and Young's modulus of the materials tested derived from the stress–strain curves are presented in Fig. 7. In determining the Young's modulus, the compliance of the testing machine, measured in separate compression tests on bulk pure Ti, was taken into account. The average values and the error bars in Fig. 7 were determined from

Fig. 5. Average pore size and average wall thickness of porous Ti/Mg and porous Ti produced by ECAP compaction of Ti, Mg and Si particles with subsequent removal of sacrificial material.

Fig. 6. Engineering stress–strain curves for porous Ti/Mg and porous Ti for different numbers of ECAP passes.

three stress–strain curves measured for each material. Clearly with 10 vol.% of Mg efficiently bonded to Ti particles, porous Ti/Mg composites with 50% porosity have higher strength and elastic modulus than porous Ti with 60% of porosity. Also, a trend of increasing strength and elastic modulus with the number of ECAP passes is seen. After 4 ECAP passes, the compressive strength and Young's modulus of porous Ti/Mg composite are 168.2 ± 7.7 MPa and 5.7 ± 0.2 GPa, respectively. A drop in the magnitude of strength and elastic modulus to, respectively, 122.3 ± 9.4 MPa and 3.9 ± 0.5 GPa after removal of Mg was recorded. This result also indicates that with this amount of Mg particles, they did not hinder the bonding between Ti particles. Therefore once they were removed the porous Ti scaffold still retained a sufficiently high strength. 3.4. Microstructure of the Ti struts in the porous structures Grain refinement evolution with increasing number of ECAP passes is presented by coloured EBSD maps in Fig. 8. A misorientation of 15° was chosen to constitute a threshold value for high angle grain boundaries (HAGB); boundaries with misorientations below 15° are considered as low angle (or subgrain) boundaries. Measurements of grain size were made directly from the EBSD images using the linear intercept method. As shown in Fig. 9, an increase in the number of ECAP

Fig. 7. Compressive strength and Young's modulus of porous Ti/Mg and porous Ti produced by ECAP compaction of Ti, Mg and Si particles with subsequent removal of sacrificial material.

Y. Qi et al. / Materials Science and Engineering C 59 (2016) 754–765

759

Fig. 8. EBSD grain orientation maps for Ti struts after ECAP processing of (a) 1 pass, (b) 2 passes, and (c) 4 passes. The white area in Fig. 8 (b) corresponds to non-indexed pixels. Black lines denote high angle grain boundaries (HAGBs).

To verify the grain size estimated from EBSD data, additional TEM measurements were carried out on specimens that underwent 4 ECAP passes; representative micrographs are shown in Fig. 10. From the bright field (BF) and dark field (DF) images, the occurrence of grains in the range from 50 to 200 nm is evident. This means at least part of the grain population can be classified as nano grains. The inset Fig. 10(c) presents a continuous ring and suggests a random misorientation and high population of high angle grain boundaries. 3.5. Characterization of morphology of pore walls

Fig. 9. Evolution of average grain size in Ti struts with increasing number of ECAP passes.

passes from 1 to 4 led to a grain refinement within the final Ti struts from 1.86 ± 1.08 μm to 789 ± 335 nm. With its average grain size in the sub-micron range, this material can be categorized as an UFG one [32].

The above results show that the porous structures produced by the methods described are tuneable and have adequate mechanical strength and ductility. However, to achieve rapid and stable osteogenesis, suitable pore surface morphology and chemistry are essential [57,58]. The pore surface topography in porous Ti is presented in SEM pictures shown in Fig. 11. After removal of Si particles, two kinds of pore surface topography were observed. One — marked by black arrows in Fig. 11(a,b) — is smooth and featureless. The other, indicated by red arrows, is rough. It is believed to be associated with the areas where the oxide layer on Ti particles was broken during deformation and after Si particles were removed, NaOH solvent reacted with fresh Ti to form titanate. The chemical reactions involved are shown below [59]:  Ti þ 3OH →TiðOHÞþ 3 þ 4e

ð3Þ

Fig. 10. TEM images of cell wall Ti subjected to 4 ECAP passes with (a) bright field image, (b) tilted dark field image, and inset (c) showing the corresponding SAD pattern using the aperture diameter of 0.75 μm.

760

Y. Qi et al. / Materials Science and Engineering C 59 (2016) 754–765

Fig. 11. Three conditions of pore surfaces in porous Ti: (a) completely smooth surface, (b) partially smooth and partially rough surfaces, and (c) completely rough portions of a surface. The inset in (c) shows surface features at high magnification.

 TiðOHÞþ 3 þ e →TiO2 H2 O þ 0:5H2 ↑

ð4Þ

Fig. 11 is based on an analysis of more than 100 pores for each condition of porous Ti, i.e. those stemming from initial compacts processed by 1, 2, and 4 ECAP passes. Statistical analysis revealed that the proportion of pores possessing a rough surface increased from 18.7% after 1 pass to 51.8% after 2 passes and then further to 73.4% after 4 passes. This trend can be explained by a higher incidence of oxide layer breakage with the increasing number of ECAP passes. The increase in the proportion of rough surfaces was accompanied with a decline in the surface roughness. A decrease of surface roughness with the number of ECAP passes is documented in Fig. 12. After Si particles were removed by NaOH 5 M at 60 °C, a porous titanate surface was formed, as shown in Fig. 12(a–c). In the case of a four-pass sample, nearly globular particles with the average size of 320 nm ± 77 nm and a smoother surface than that after one and two ECAP passes were observed. To remove Mg, 100 mM HCl was used and the surface morphology was changed. Topography features in Fig. 12(f) have the smallest scale and correspond to the smoothest pore surface. Because these features were located on the pore surfaces, we could not conduct surface topography measurements by atomic force microscopy. Nevertheless, from SEM characterization the trend of decreasing roughness with increasing number of

ECAP passes is clear. It is also worth noting that this decreased pore surface roughness is believed to be an important factor in the enhanced strength and stiffness of the porous Ti [60].

3.6. Cell viability Since the fabrication route proposed is new, it is necessary to do a preliminary cell viability test to verify the biocompatibility of the porous Ti fabricated using this processing recipe. As shown in Fig. 13, based on the measurements of the metabolic activity, in short-term (3 days) there was no significant difference in terms of the MC3T3-E1 preosteoblast cell proliferation between porous and solid Ti, which indicates that viable cells and initial attachment did not depend much on the structure during the three-day incubation. However, 4 days later the number of viable cells on porous four-pass Ti was increased substantially and was higher than for one- and two-pass porous Ti or bulk Ti. This result suggests that, based on 7 days incubation, the pore surface morphology of porous Ti that went through four ECAP passes at the compaction stage was most favourable for cell proliferation. Since it is generally accepted that micro/nanotopology benefits the proliferation of MC3T3-E1 preosteoblast cells [61], this higher cell viability can be

Fig. 12. Representative SEM images of surfaces of pore walls in Ti (a–c) after removal of Si particles by 5 M NaOH, (d–f) after subsequent removal of Mg particles by 100 mM HCl. Note the evolution of the surface features with increasing number of ECAP passes.

Y. Qi et al. / Materials Science and Engineering C 59 (2016) 754–765

Fig. 13. Cell viability measured by MTS assay after culturing for 3 and 7 days on the porous Ti scaffolds from 1, 2, 4 ECAP processes.

presumably explained by the largest surface area with micro/ nanotopography features in the four-pass material. Despite the difference of the roughness of the porous titanate surface in the materials that experienced different numbers of ECAP passes, cells were attached to the inclined porous titanate surfaces, Fig. 14(a–c). With higher magnification shown in Fig. 14(d–f), the lamellipodia and filopodia were developed and interacted with the porous titanate substrate manifested as membrane tethers. The cell-substrate interaction observed in SEM images suggests that the surfaces produced by the present processing sequence are beneficial for cell attachment. 4. Discussion 4.1. Redistribution of the constituents during ECAP The observed increase of pore size and wall thickness with the increasing number of ECAP passes during the compaction of the elemental powder blends is a somewhat unexpected result. It can be rationalized in terms of a redistribution of Ti, Mg and Si during ECAP processing. The distribution of the different elements in the green compacts of Ti/Mg/Si powder consolidated by ECAP is illustrated by

761

backscattered electron (BSE) images based on the atomic number (Z) contrast, cf. Fig. 15(a–c). It is seen that Ti particles were heavily deformed and merged with neighbouring Ti, Mg, or Si particles after the first pass. No fragmentation of Si particles is seen in Fig. 15(d,g). A fraction of Mg-rich regions (Mg particles) have an elongated shape, which indicates that they went through heavy shear deformation, Fig. 15(j). However, the rest of Mg regions retained their original shapes, as one region is marked by arrow in Fig. 15(j). When the composite underwent further ECAP passes, Si particles tended to fracture, whilst Mg and Ti particles as softer constituents were ‘extruded’ into the voids created by the fracturing. In the samples with four-pass history almost all Mg constituents were elongated and located predominantly around Si particles, Fig. 15(l). Some Si particles are delineated by dashed lines in Fig. 15(f,i,l). From these images it can also be surmised that clustering or agglomeration of Si took place. Si agglomeration can also be found in samples after two ECAP passes, yet to a lesser extent, Fig. 15(b,e,h). Almost no clustering was observed for the single ECAP-pass material, Fig. 15(d,j). Redistribution of particles in a composite during processing is not a new finding [62]. However, in this study, due to a mismatch of Young's modulus and particle size of the constituent particles, redistribution caused segregation of Ti and Si particles yet resulted in a more uniform distribution and size refinement of Mg particles. The level of Ti and Si particle agglomeration produced by ECAP processing is regarded as beneficial, since it promoted larger pore size and stronger struts without having a detrimental effect on porosity and pore interconnectivity. 4.2. ECAP-assisted bonding between the constituents One of the merits of ECAP-assisted compaction that makes it stand out from other powder consolidation techniques is its ability to achieve strong bonding at relatively low temperature via fresh metal contact and solid state interdiffusion. Imposed severe plastic deformation of the particles, and especially severe shear promoting breakage of oxide layers, is essential for creating fresh metal surface contact. In Fig. 16, an interface between two Ti particles in a Ti/Mg/Si compact produced by four ECAP passes is presented in STEM-BF and high-angle annular dark field (HAADF) images. An object with a low atomic number perpendicular to the interface is seen in Fig. 16(a,b). With higher magnification Fig. 16 (c, d) and the corresponding EDX mappings Fig. 16 (e, f), this object can be identified as a pair of Ti oxide particles with 10 nm thickness stacked or folded together. The rest of the interface is free of

Fig. 14. SEM images of preosteoblast MC3T3-E1 cells cultured for 3 days on porous Ti 1p, 2p and 4p at the magnification of 2500× (a–c) and the magnification of 10,000× with details of filopodia (d–f).

762

Y. Qi et al. / Materials Science and Engineering C 59 (2016) 754–765

Fig. 15. SEM-BSE images (a–c) and SEM-EDX mappings (d–l) of Ti/Mg/Si compacts after 1, 2 and 4 ECAP passes.

oxide layer. From our previous study [53] and the electron microscopy data obtained in the present work it can be conjectured that with the growing number of ECAP passes and the greater shear deformation imparted to powder particles, greater damage to the oxide layer is induced and fresh metal exposure is increased. High pressure involved in ECAP processing is also conducive for ‘fusion’ between the adjacent particles stripped bare of the oxide layers. The highest strength of porous Ti that went through 4 ECAP passes in its compaction history is believed to be caused by this mechanism. Strong bonding between Ti and Mg is also important, since without effective bonding Mg particles would not contribute much to the strength of porous Ti/Mg composites. Fig. 17(a) and (b) present BF and HAADF images of interfaces between Ti and Mg particles. Diffuse boundaries between the constituent particles indicate that migration of Ti and Mg atoms across the interfaces occurred promoting bonding by interdiffusion. From Fig. 17(d), it is clear that no oxygen-enriched region exists at the interface, which indicates that both Ti and Mg oxide layers were effectively broken during ECAP processing thus promoting bonding.

5. Conclusions Compaction by ECAP was utilized to consolidate an elemental powder blend, which comprised Ti, Mg and Si powders. Readily machinable compacts were produced in this way. By removing Si and the Mg constituents from the compact using chemical reactions, porous Ti and porous Ti/Mg composite were obtained. Due to the advantages provided by ECAP compaction, including the relatively low process temperature, the porous structures produced possessed unique properties. First, ECAP compaction resulted in exceptional grain refinement in Ti struts down to sub-micron range after four ECAP passes. This grain refinement gave rise to high compressive strength and acceptable ductility. Second, ECAP processing led to a redistribution of the constituents, leading to thicker struts and larger average pore size with increasing number of ECAP passes. With broader strut thickness their strength increased as well. Third, ECAP deformation led to breakage of the oxide layer of Ti and Mg particles thus promoting solid state interdiffusion between them. The quality of bonding between titanium particles as well as titanium and magnesium particles was improved in this way, as were

Y. Qi et al. / Materials Science and Engineering C 59 (2016) 754–765

763

Fig. 16. Interface between Ti particles in a compact obtained by 4 ECAP passes: (a) STEM-BF image and (b) STEM-HAADF image of interface between two Ti particles ‘fused’ together. Crosssections (c) and (d) show an oxide fragment embedded in Ti; (e) and (f) are the corresponding STEM-EDX mappings showing the distribution of titanium and oxygen, respectively.

Fig. 17. Interfaces between Ti and Mg particles in a compact after 4 ECAP passes: (a) STEM-BF and (b) STEM-HAADF image of interfaces between Ti and Mg; (c) and (d) present the corresponding STEM-EDX mappings.

764

Y. Qi et al. / Materials Science and Engineering C 59 (2016) 754–765

strength and ductility of the struts. Also, a unique morphology of pore surfaces was obtained as a result of a chemical reaction between NaOH and fresh Ti. This pore surface morphology was found to be conducive for interaction of the porous scaffolds with preosteoblast cells. Acknowledgements The authors gratefully acknowledge the excellent technical assistance provided by Monash Centre of Electron Microscopy (MCEM) and Monash Micro Imaging (MMI). YQ is grateful to Dr. Xiaobo Chen for technical discussions. Funding support from the Russian Ministry for Education and Science (grant #14.A12.31.0001) is highly appreciated. References [1] J.J. Klawitter, J.G. Bagwell, A.M. Weinstein, B.W. Sauer, J.R. Pruitt, An evaluation of bone growth into porous high density polyethylene, J. Biomed. Mater. Res. 10 (1976) 311–323. [2] R. Pilliar, H. Cameron, I. Macnab, Porous surface layered prosthetic devices, Biomed. Eng. 10 (1975) 126–131. [3] P. Ducheyne, P. De Meester, E. Aernoudt, M. Martens, J.C. Mulier, Influence of a functional dynamic loading on bone ingrowth into surface pores of orthopedic implants, J. Biomed. Mater. Res. 11 (1977) 811–838. [4] J. Bobyn, R. Pilliar, H. Cameron, G. Weatherly, The optimum pore size for the fixation of porous surfaced metal implants by the ingrowth of bone, Clin. Orthop. Relat. Res. 150 (1980) 263–270. [5] C. Engh, J. Bobyn, A. Glassman, Porous-coated hip replacement. The factors governing bone ingrowth, stress shielding, and clinical results, J. Bone Joint Surg. 69 (1987) 45–55. [6] H. Kröger, P. Venesmaa, J. Jurvelin, H. Miettinen, O. Suomalainen, E. Alhava, Bone density at the proximal femur after total hip arthroplasty, Clin. Orthop. Relat. Res. 352 (1998) 66–74. [7] R. Huiskes, H. Weinans, B. van Rietbergen, The relationship between stress shielding and bone resorption around total hip stems and the effects of flexible materials, Clin. Orthop. Relat. Res. 274 (1992) 124–134. [8] D. Kuroda, M. Niinomi, M. Morinaga, Y. Kato, T. Yashiro, Design and mechanical properties of new β type titanium alloys for implant materials, Mater. Sci. Eng. A 243 (1998) 244–249. [9] M. Geetha, A. Singh, R. Asokamani, A. Gogia, Ti based biomaterials, the ultimate choice for orthopaedic implants—a review, Prog. Mater. Sci. 54 (2009) 397–425. [10] J.A. Kanis, L.J. Melton, C. Christiansen, C.C. Johnston, N. Khaltaev, The diagnosis of osteoporosis, J. Bone Miner. Res. 9 (1994) 1137–1141. [11] S. Wu, X. Liu, K.W.K. Yeung, C. Liu, X. Yang, Biomimetic porous scaffolds for bone tissue engineering, Mater. Sci. Eng. R 80 (2014) 1–36. [12] M. Long, H. Rack, Titanium alloys in total joint replacement—a materials science perspective, Biomaterials 19 (1998) 1621–1639. [13] H. Rack, J. Qazi, Titanium alloys for biomedical applications, Mater. Sci. Eng. C 26 (2006) 1269–1277. [14] Y. Li, C. Yang, H. Zhao, S. Qu, X. Li, Y. Li, New developments of Ti-based alloys for biomedical applications, Materials 7 (2014) 1709–1800. [15] M.T. Choy, C.Y. Tang, L. Chen, C.T. Wong, C.P. Tsui, Mater. Sci. Eng. C 42 (2014) 746–756. [16] G. Ryan, A. Pandit, D.P. Apatsidis, Fabrication methods of porous metals for use in orthopaedic applications, Biomaterials 27 (2006) 2651–2670. [17] S.J. Hollister, Scaffold design and manufacturing: from concept to clinic, Adv. Mater. 21 (2009) 3330–3342. [18] R. Huiskes, The various stress patterns of press-fit, ingrown, and cemented femoral stems, Clin. Orthop. Relat. Res. 261 (1990) 27–38. [19] V. Karageorgiou, D. Kaplan, Porosity of 3D biomaterial scaffolds and osteogenesis, Biomaterials 26 (2005) 5474–5491. [20] S. Van Bael, Y.C. Chai, S. Truscello, M. Moesen, G. Kerckhofs, H. Van Oosterwyck, J.P. Kruth, J. Schrooten, The effect of pore geometry on the in vitro biological behavior of human periosteum-derived cells seeded on selective laser-melted Ti6Al4V bone scaffolds, Acta Biomater. 8 (2012) 2824–2834. [21] L.J. Gibson, M.F. Ashby, B.A. Harley, Structure and Mechanics of Cellular Materials, Cellular Materials in Nature and Medicine, Cambridge University Press, 2010 46–47. [22] C.Y. Lin, T. Wirtz, F. LaMarca, S.J. Hollister, Structural and mechanical evaluations of a topology optimized titanium interbody fusion cage fabricated by selective laser melting process, J. Biomed. Mater. Res. A 83 (2007) 272–279. [23] D.K. Pattanayak, A. Fukuda, T. Matsushita, M. Takemoto, S. Fujibayashi, K. Sasaki, N. Nishida, T. Nakamura, T. Kokubo, Bioactive Ti metal analogous to human cancellous bone: fabrication by selective laser melting and chemical treatments, Acta Biomater. 7 (2011) 1398–1406. [24] A. Fukuda, M. Takemoto, T. Saito, S. Fujibayashi, M. Neo, D.K. Pattanayak, T. Matsushita, K. Sasaki, N. Nishida, T. Kokubo, T. Nakamura, Osteoinduction of porous Ti implants with a channel structure fabricated by selective laser melting, Acta Biomater. 7 (2011) 2327–2336. [25] B. Vrancken, L. Thijs, J.-P. Kruth, J. Van Humbeeck, Heat treatment of Ti6Al4V produced by selective laser melting: microstructure and mechanical properties, J. Alloys Compd. 541 (2012) 177–185.

[26] K. Kato, S. Ochiai, A. Yamamoto, Y. Daigo, K. Honma, S. Matano, K. Omori, Novel multilayer Ti foam with cortical bone strength and cytocompatibility, Acta Biomater. 9 (2013) 5802–5809. [27] S.-W. Yook, H.-D. Jung, C.-H. Park, K.-H. Shin, Y.-H. Koh, Y. Estrin, H.-E. Kim, Reverse freeze casting: a new method for fabricating highly porous titanium scaffolds with aligned large pores, Acta Biomater. 8 (2012) 2401–2410. [28] C.E. Wen, M. Mabuchi, Y. Yamada, K. Shimojima, Y. Chino, T. Asahina, Processing of biocompatible porous Ti and Mg, Scr. Mater. 45 (2001) 1147–1153. [29] B. Arifvianto, J. Zhou, Fabrication of metallic biomedical scaffolds with the space holder method: a review, Materials 7 (2014) 3588–3622. [30] T. Watanabe, Y. Horikoshi, The sintering phenomenon of titanium powders—a discussion, Int. J. Powder Metall. Powder Technol. 12 (1976) 209–214. [31] G.B. Schaffer, M. Qian, C.J. Bettles, Sintering of titanium and its alloys, in: Z.Z. Fang (Ed.), Sintering of Advanced Materials, Woodhead Publishing 2010, pp. 336–338. [32] Y. Estrin, A. Vinogradov, Extreme grain refinement by severe plastic deformation: a wealth of challenging science, Acta Mater. 61 (2013) 782–817. [33] A.Y. Vinogradov, V.V. Stolyarov, S. Hashimoto, R.Z. Valiev, Cyclic behavior of ultrafine-grain titanium produced by severe plastic deformation, Mater. Sci. Eng. A 318 (2001) 163–173. [34] X. Zhao, X. Yang, X. Liu, X. Wang, T.G. Langdon, The processing of pure titanium through multiple passes of ECAP at room temperature, Mater. Sci. Eng. A 527 (2010) 6335–6339. [35] Y. Okazaki, S. Rao, T. Tateishi, Y. Ito, Cytocompatibility of various metal and development of new titanium alloys for medical implants, Mater. Sci. Eng. A 243 (1998) 250–256. [36] A. Medvedev, H. Ng, R. Lapovok, Y. Estrin, T. Lowe, V. Anumalasetty, Comparison of laboratory-scale and industrial-scale equal channel angular pressing of commercial purity titanium, Mater. Lett. 145 (2015) 308–311. [37] T.G. Langdon, Twenty-five years of ultrafine-grained materials: achieving exceptional properties through grain refinement, Acta Mater. 61 (2013) 7035–7059. [38] L. Mishnaevsky Jr., E. Levashov, R.Z. Valiev, J. Segurado, I. Sabirov, N. Enikeev, S. Prokoshkin, A.V. Solov'yov, A. Korotitskiy, E. Gutmanas, I. Gotman, E. Rabkin, S. Psakh'e, L. Dluhoš, M. Seefeldt, A. Smolin, Nanostructured titanium-based materials for medical implants: modeling and development, Mater. Sci. Eng. R 81 (2014) 1–19. [39] Y. Estrin, E.P. Ivanova, A. Michalska, V.K. Truong, R. Lapovok, R. Boyd, Accelerated stem cell attachment to ultrafine grained titanium, Acta Biomater. 7 (2011) 900–906. [40] Y. Estrin, C. Kasper, S. Diederichs, R. Lapovok, Accelerated growth of preosteoblastic cells on ultrafine grained titanium, J. Biomed. Mater. Res. A 90 (2009) 1239–1242. [41] J.W. Park, Y.J. Kim, C.H. Park, D.H. Lee, Y.G. Ko, J.H. Jang, C.S. Lee, Enhanced osteoblast response to an equal channel angular pressing-processed pure titanium substrate with microrough surface topography, Acta Biomater. 5 (2009) 3272–3280. [42] T.N. Kim, A. Balakrishnan, B. Lee, W. Kim, B. Dvorankova, K. Smetana, J. Park, B. Panigrahi, In vitro fibroblast response to ultra fine grained titanium produced by a severe plastic deformation process, Mater. Sci. Mater. Med. 19 (2008) 553–557. [43] F. Nie, Y. Zheng, S. Wei, D. Wang, Z. Yu, G. Salimgareeva, A. Polyakov, R. Valiev, In vitro and in vivo studies on nanocrystalline Ti fabricated by equal channel angular pressing with microcrystalline CP Ti as control, J. Biomed. Mater. Res. A 101 (2013) 1694–1707. [44] S. Bagherifard, R. Ghelichi, A. Khademhosseini, M. Guagliano, Cell response to nanocrystallized metallic substrates obtained through severe plastic deformation, ACS Appl. Mater. Interfaces 6 (2014) 7963–7985. [45] T.C. Lowe, R.A. Reiss, Understanding the biological responses of nanostructured metals and surfaces, IOP Conf. Ser. Mater. Sci. Eng. 63 (2014) 012172–012189. [46] R.Z. Valiev, T.G. Langdon, Principles of equal-channel angular pressing as a processing tool for grain refinement, Prog. Mater. Sci. 51 (2006) 881–981. [47] R. Lapovok, D. Tomus, C. Bettles, Shear deformation with imposed hydrostatic pressure for enhanced compaction of powder, Scr. Mater. 58 (2008) 898–901. [48] R. Lapovok, D. Tomus, B.C. Muddle, Low-temperature compaction of Ti–6Al–4V powder using equal channel angular extrusion with back pressure, Mater. Sci. Eng. A 490 (2008) 171–180. [49] J.R. Henstock, L.T. Canham, S.I. Anderson, Silicon: the evolution of its use in biomaterials, Acta Biomater. 11 (2015) 17–26. [50] F. Witte, The history of biodegradable magnesium implants: a review, Acta Biomater. 6 (2010) 1680–1692. [51] Z. Esen, B. Dikici, O. Duygulu, A.F. Dericioglu, Titanium–magnesium based composites: mechanical properties and in vitro corrosion response in ringer's solution, Mater. Sci. Eng. A 573 (2013) 119–126. [52] G. Jiang, C. Wang, Q. Li, J. Dong, G. He, Porous titanium with entangled structure filled with biodegradable magnesium for potential biomedical applications, Mater. Sci. Eng. C 47 (2015) 142–149. [53] R. Lapovok, Y. Qi, H. Ng, V. Maier, Y. Estrin, Multicomponent materials from machining chips compacted by equal-channel angular pressing, J. Mater. Sci. 49 (2014) 1193–1204. [54] S.W. Kim, H.-D. Jung, M.-H. Kang, H.-E. Kim, Y.-H. Koh, Y. Estrin, Fabrication of porous titanium scaffold with controlled porous structure and net-shape using magnesium as spacer, Mater. Sci. Eng. C 33 (2013) 2808–2815. [55] D.K. Pattanayak, S. Yamaguchi, T. Matsushita, T. Nakamura, T. Kokubo, Apatiteforming ability of titanium in terms of pH of the exposed solution, J. R. Soc. Interface 9 (2012) 2145–2155. [56] J.C. Li, D.C. Dunand, Mechanical properties of directionally freeze-cast titanium foams, Acta Mater. 59 (2011) 146–158. [57] S. Amin Yavari, J. van der Stok, Y.C. Chai, R. Wauthle, Z. Tahmasebi Birgani, P. Habibovic, M. Mulier, J. Schrooten, H. Weinans, A.A. Zadpoor, Bone regeneration performance of surface-treated porous titanium, Biomaterials 35 (2014) 6172–6181.

Y. Qi et al. / Materials Science and Engineering C 59 (2016) 754–765 [58] S. Amin Yavari, Y.C. Chai, A.J. Böttger, R. Wauthle, J. Schrooten, H. Weinans, A.A. Zadpoor, Effects of anodizing parameters and heat treatment on nanotopographical features, bioactivity, and cell culture response of additively manufactured porous titanium, Mater. Sci. Eng. C 51 (2015) 132–138. [59] H.M. Kim, F. Miyaji, T. Kokubo, T. Nakamura, Preparation of bioactive Ti and its alloys via simple chemical surface treatment, J. Biomed. Mater. Res. 32 (1996) 409–417. [60] N. Tuncer, G. Arslan, E. Maire, L. Salvo, Investigation of spacer size effect on architecture and mechanical properties of porous titanium, Mater. Sci. Eng. A 530 (2011) 633–642.

765

[61] X. Zhuang, B. Zhou, J. Ouyang, H. Sun, Y. Wu, Q. Liu, F. Deng, Enhanced MC3T3-E1 preosteoblast response and bone formation on the addition of nano-needle and nano-porous features to microtopographical titanium surfaces, Biomed. Mater. 9 (2014) 045001–045012. [62] J. Park, D.X. Lenshek, G.L. Povirk, Reinforcement redistribution in Al–SiC composites under cyclic deformations, Acta Mater. 45 (1997) 1351–1364.